Effects of grassland responses to elevated atmospheric carbon dioxide on evapotranspiration and recharge

The elevated concentration of atmospheric carbon dioxide (eCO₂) resulting from anthropogenic emissions has led to increasing temperatures and thus increasing evaporative demand of the atmosphere. The increased evaporative demand appears to suggest increased evapotranspiration (ET) and thus decreasing groundwater recharge. Yet, observational data and climate-model outputs reveal a more complex pattern of ET responses to climate change. In particular, it has been suggested that the commonly used Penman-Monteith equation overpredicts changes in potential evapotranspiration, partly because the effect of eCO₂ on stomatal resistance is neglected (Milly & Dunne 2016). Here we aim (i) to identify the effects of eCO₂ on ET using experimental data from a managed alpine grassland, (ii) to quantify the effect of eCO₂ on the stomatal resistance of the managed alpine grassland, (iii) to assess the errors in simulated soil-water fluxes and thus groundwater recharge caused by ignoring stomatal effects of eCO₂.

To identify effects of eCO_2, we analyze data from two weighable high-precision lysimeters that are part of the ClimGrass climate change experiment operated by the Agricultural Research and Education Centre Raumberg-Gumpenstein (Austria). One of the lysimeters is treated using the Free Air Carbon dioxide Enrichment (FACE) technique to achieve a controlled increase of the atmospheric carbon dioxide (+ 300 ppm) relative to the ambient conditions, under which the other (reference) lysimeter is operated. The lysimeter fluxes reveal a lower ET of the treated (eCO₂) lysimeter throughout the observation period. A remarkable exception, however, is found during two dry periods in 2018 and 2019. In those periods, ET was found to be limited by the depleted soil moisture at the untreated lysimeter, whereas at the treated lysimeter, the stomatal effect of eCO₂ enabled water savings and thus higher ET during the dry periods.

The stomatal resistance was estimated for each of the two lysimeters using an inverse modelling approach based on meteorological and phenological data, as well as the lysimeter ET observed on days without water stress. A stomatal resistance of 60 s/m and 90 s/m was found for the untreated and treated lysimeter, respectively. Further, a relative sensitivity of stomatal resistance to eCO₂ was obtained, which can be used to obtain estimates of stomatal resistance for other levels of eCO₂. It is noteworthy that these results refer to the managed alpine grassland at the experimental site and cannot directly be generalized. Still, the findings agree reasonably well with studies on other grasslands.

Hydrological modelling of the soil-water fluxes in the lysimeters reveals that neglecting the effect of eCO₂ in the given case causes an underestimation of annual recharge by up to 5%. Accounting for the eCO₂ stomatal effect particularly improves the simulation of recharge events after dry periods.